By Peter Branton
Previously thought of as a long-term strategic tool, business intelligence (BI) has moved onto the shop floor to help managers deal swiftly with everyday activities
Introduction|~||~||~|Business intelligence (BI) vendors are bringing their solutions to the masses — that is, moving them out of the boardroom and into the front line. No longer just for senior executives and other key decision makers, BI is now being simplified to adapt to the requirements of an enterprise’s operational staff.
Previously, BI was viewed as a strategic and tactical reporting and analysis tool for top management and business analysts who tend to look at business trends from a broader perspective.
BI in this space would support those people who manage their business by providing an overall view of their organisation and helping them pinpoint certain areas for improvement.
At this level, there are long decision-making cycles, which can take several weeks or months, and rely on aggregated data with several months or years of history. The emergence of a new form of BI, called operational business intelligence, deals more with the day-to-day activities of an organisation and is geared for a more general audience: the company’s employees.
By bringing BI down to this level, a salesperson or a customer representative, for instance, is given a tool that can help them make informed decisions on the spot — whether they are on the phone with a client or on the road making a sales visit.
Lisa Dreyer, director of product marketing manager at Sybase, calls this new application of BI as the “right-time” approach. “Right-time BI is all about delivering the right information in the right format to the right people at the right time for decision-making purposes,” she says.
“Operational BI came about as a new way of BI in allowing BI to become more pervasive and deliver more value by being further entrenched into the business, so that decisions made on a day-to-day basis are supported by BI rather than some of the more strategic and tactical approaches that BI has been applied to today,” says Ian Parker, product and solution marketing manager, Business Objects Asia Pacific.
“Currently BI is used as a strategic tool in its dashboarding, scorecarding and performance management capabilities and it is seen as a tactical tool in that it can be used to monitor and analyse a business activity over time and analyse how that can be improved and changed, so taking a very historical view at a time. What process-aware BI is about is taking BI down into the day-to-day activities of a business and linking it into the operations,” he adds.
The goal of operational BI, primarily, is to shorten the latency between when a business event (or process) takes place and an action (or decision) is made. This latency normally consists of three elements: the time it takes to get the data ready for analysis, the time it takes to analyse the data and generate the results, and the time it takes for the person to receive the results of the analysis and understand what action must be taken. For operational BI to be effective, these three latencies should be minimised to virtually zero.
Operational BI, according to David Brierly, regional manager of Cognos Middle East, is not a new concept. In fact, companies have been doing operational reporting for many years.
The emergence of operational BI, he thinks, stems from the fact that previous solutions were not capable of handling the three levels of BI: strategic, tactical and operational.
“There weren’t many BI vendors that could address all three of those requirements, which means that if you wanted to have strategic dashboarding you can go to the market and there will be ten players but none of those players can help you with production reporting,” Brierly says.
“Due to the limitations of the technology, organisations happen to purchase really different products which cause problems,” he continues.
“The major problem it caused is on the metadata layer. With the operational systems, they have lots of data. Each operational system is capturing its own data. With separate products you find that simple things like interpreting a customer [tends to be] different,” he adds.
Because of this inconsistency in data and process definition, users tend to get different
results from different operational systems.
“The business logic is different because each department may view customer differently,” Brierly says.
“To realise the full benefits of BI you must have a clear definable business layer where the actual process definition have been agreed and agreed across the whole enterprise,” he adds.
||**||One single truth|~||~||~|From a technical standpoint, Brierly claims that companies like Cognos and Business Objects can clearly now support the requirements of all three layers of BI. He also observes that companies are working with solution providers to identify their unique process definitions and introduce them in their BI tools.
“Companies are now working on a single version of truth because everyone is working on the same process definitions. Two to three years ago you could never have had that because the technology couldn’t support what unique process definitions are and couldn’t do all of what is required,” Brierly states. “The two major vendors — Cognos and Business Objects — can do that now. It’s not a problem any more,” he adds.
Asif Chhapra, professional services manager, Hyperion Solutions Middle East, agrees that operational BI is nothing new.
It has emerged as a fundamental requirement but he claims it has been there since ERP came to the fore in the late 1990s.
“It is only through recent technological advancements that BI has become completely stable, where this technology actually provides the long-awaited demand of large organisations. If you look at operational BI, BI itself is divided into two: advanced analytics, which is three-dimensional online analytical processing (OLAP), and the BI for
relational reporting, under which operational BI comes into,” Chhapra says.
According to Ventana Research analyst Dan Everett, while BI vendors have talked about operational BI or “BI for the masses” Ventana’s research has shown that firms, so far, have deployed their scorecards, dashboards and alerting tools to decision makers that are relatively high up in the organisation.
In a report, Everett discovers that the use of BI by executives and mid-level managers is more than double the use by business analysts and more than triple the use by front-line employees.
“We also have found that companies use BI primarily to manage strategic directions and the financial health of the business rather than to make operational decisions. Use in finance and sales, for example, is nearly triple that of the supply chain or customer operations,” Everett wrote in the report.
So, what has brought this growing emphasis to put BI at the lowest level — the operational level — of an organisation?
Competitive pressure, mainly. In order to stay ahead in the market, companies are forced to react quicker to changing business circumstances and customer demands.
New compliance policies are also pressuring organisations to perform better, as they are required to be more transparent about their wins and — most especially — their losses. Customer expectations are becoming much higher and margins for errors are shrinking.
The goal of operational BI is to tightly link analytical applications to business processes in order to ensure that right at the operational level, employees are provided with real-time data that can guide them to make faster informed decisions.
“The idea is to take BI down into the day-to-day lives of people and make it embedded into the processes that they work with everyday and help them make the decisions that they make everyday,” explains Parker.
“It’s about taking people to a decision-centric BI, aligning BI with the decisions that people make everyday, and many of the decisions they make are process-centric,” he adds.
“Operational BI is a BI system exclusively focusing on areas, which include decision making and information dissemination at the operational level by different line staff. Operational BI basically automates and accelerates their business processes and
decision making, delivering information directly to the point of business and reducing the cost of doing business,” explains Chhapra. “It is not only exclusively for the frontline staff but operational BI can also be extended to business partners and customers to solve problems fast and stay ahead in a volatile market,” he adds.
With operational BI, employees can have access to interactive reports that are based on operational data, which they can use to make split-second decisions about what should be done immediately. It can help them play “what if” scenarios in real time and embed corrective action into business processes. Also, it can help them monitor business activities as they happen. Operational BI can improve the activities and processes of thousands of workers by supporting work in a more streamlined and efficient manner.
By deploying operational BI, companies can prevent minor mistakes from turning into corporate disasters.
There are various applications for operational BI. In the retail segment, for example, a grocery store can use it to help its cashier identify the individual tastes of your customers.
Having access to real-time information, the cashier can review the client’s purchasing history, cross reference that to the grocery’s weekly promotion and be able to make some helpful suggestions. That would not only possibly lead to higher sales but also improved customer satisfaction and loyalty.
Another potential use of operational BI can be found in the airline industry, according to Chhapra. “For example, an airline company, which has a strategic partnership with a large conglomerate, is negotiating to provide 5000 return tickets to the Far East for its employees for a period of one year. With operational BI, the airline sales executive can instantaneously change the end user experience completely if he is able to provide the client with relevant information, such as costing and the margins of the people on the spot itself,” Chhapra says.
By providing the information in what Chhapra refers to as the “speed of thought” the customer will be able to decide right there and then instead of delaying the process in two or five days.
He adds that customer-facing units of an organisation are most likely to benefit more from operational BI.
“Industries like the airline, automotive, pharmaceutical and healthcare are ideally placed to have operational BI. All these industries have very large customer interactions required. These are the industries that can really benefit from operational BI,” Chhapra says.
Implementing operational intelligence, however, is challenging as it involves not only technology but more importantly, people and knowledge.
Information sharing, for instance, is critical as operational BI would require everyone involved in the process to have full access to information to be able to change the ways that they work.
“Information sharing is the most important, the kind of information that comes up, the quality of it, the scalability and the way that it is disseminated. It is very, very important at the operational level,” Chhapra says.
Basel Tutunji, regional manager, SAS Middle East, says companies should look at different aspects of data.
“One of the components that they should consider is data integration. This is a major issue currently in several organisations. How do you locate data? Where do you find data? And how do you integrate data? This is the major pain when it comes to BI,” he continues.
“We do realise that a lot of customers are looking into data integration as the key point before even walking to full BI,” Tutunji explains. “They also have serious gaps when it comes to availability of data and the quality of the data and the unification of the data,” he adds.
As there is a higher turnaround time on the front lines, this could also affect the decisions made in the operational level.
Not only will it take time to hire and train new employees about the best practices that the company uses on a daily basis but he or she will have to get used to making the right kind of decisions that can only be achieved through experience and looking at the right kind of information.
For example, a new hire in inventory may get a notification that stock on hand has dropped below the reorder threshold. His or her reaction may be to reorder based on which vendor can supply the part the cheapest or soonest.
In fact, though, it may be better to correlate the below-threshold condition with other information, such as current orders, production schedules or in-transit logistics.
Thus, a more experienced inventory worker might run a report on forecasted demand to analyse whether price or lead time should be the primary metric to use in deciding which vendor to order from.
Compared to traditional BI, which depends on historical data, operational BI relies more on real-time transactional data.
It is necessary for companies that are looking at implementing operational BI to provide its users access to real-time operational data and access to real-time decision making.
||**||Analytical tools|~||~||~|“One of the differences between operational BI and strategic and tactical BI is that operational BI really relies on operational data. It’s not a data warehouse focused approach. It’s much more about using the operational data itself. So, it is one of the changes and that is linked to the real-time need as much as anything,” Parker says.
Operational BI also lacks the analytical skill sets needed by traditional BI, according to Parker.
“In terms of skill sets, there are no complex analytical requirements for operational BI. For operational BI to be successful analytics need to be embedded into the day-to-day activities that a person undertakes, and so it needs to be seamless,” Parker explains.
“They don’t need to have analytic capabilities. What they should be provided with is analytics that are embedded into their day-to-day processes that provide them with the information they need to make a decision. They should have some of the complexities removed from that by the support of good BI and good analytics for the particular need that they have,” he adds.
“That is what part of the job needs to do is to take operational data and put it through a statistical processing capability and then provide the output of that to the end users so that they can make better decisions,” he goes on to say.
And unlike traditional BI, which links itself with data warehouses or data marts, operational BI should be an integral part of an online transaction processing (OLTP) application, which comes in two parts.
The traditional OLTP component refers to the automation of the business process, such as an airline reservation process, which would cover customer data collection, flight scheduling, payment and ticket issuance.
The operational BI component of OLTP includes embedded business logic that would
play out at pre-defined business scenarios.
In our airline ticketing example, that would involve things like identifying underperforming routes and devising ways to improve these areas, such as offering discounts to customers.
Such a facility should be available to the airline reservation agents or to passengers booking flights online.
“You need to look at embedding BI. One of the things they need to look for is the capability of the BI solution to embed the analytics into the day-to-day operational systems that their people are using. That is one key requirement,” suggests Parker.
“They also need, within that, the ability to have an alerting capability to be able to alert people when an operational process or activity is occurring outside of acceptable limits so they can make operational decisions based on that in near real time,” he states.
“The other thing that they need on top of their standard BI system is a concept of framework for defining best practices and guiding users through that process in a repeatable way,” he continues.
“In other words, you are creating repeatable decision-making process by capturing best practices, which will basically increase your proficiency in making decisions and making them quicker, more effective and operational, and more accurate,” Parker goes on to say.
Before an organisation begins implementing operational BI, Ventana’s Everett, in his report, suggests that companies should start documenting the business rules and best practices for operational decision-making in demand, supply and financial processes.
He recommends that training documents should be provided to new employees, which, at
the same time, they can use to define the intelligence they need to embed into the operational processes.
It is also important that they assess the technology components they are currently deploying to ensure that these are capable of supporting their operational decision-making needs.
“In the near term, companies should also evaluate their current BI vendor(s) in terms of their plans for providing a service-oriented architecture that facilitates integration of traditional query, reporting and analysis capabilities with the event-based technologies required for operational intelligence,” Everett states.
“Over time, look for innovative vendors that take the required technology components from different disciplines and integrate them into a comprehensive solution that addresses the needs of both BI and operational BI,” he adds.
In general, however, the ground rules for BI have not changed. BI is still about being able to make better decisions by analysing variables, predicting trends and delivering information. Operational BI can do this and more. By making BI both more meaningful to more users, a tighter alignment between IT and the business can be developed.